Time-of-use electricity price time period division method and device, electronic equipment and computer readable storage medium

Abstract

The application discloses a time-of-use electricity price period division method, comprising: acquiring a power load sample, wherein the power load sample comprises power load data collected in a preset time length according to a preset time interval; determining initial clustering parameters of a Gaussian mixture model and a preset clustering number, wherein the preset clustering number is a period division number for the preset time length; performing parameter estimation on the initial clustering parameters according to the power load sample and the preset clustering number to obtain optimal clustering parameters; and determining a division period to which a collection time point corresponding to each power load data belongs according to the optimal clustering parameters. The technical solution provided by the application can realize more accurate and reasonable division of time-of-use electricity price. The application also discloses a time-of-use electricity price period division device, an electronic device and a computer readable storage medium, which also have the above technical effects.

Description

Technical Field

[0001] This application relates to the field of electricity market technology, and in particular to a time-of-use pricing method for dividing time periods, as well as a time-of-use pricing device, electronic equipment, and computer-readable storage medium. Background Technology

[0002] In recent years, with the rapid and sustained development of the economy, the power industry, as an important pillar of the national economy, has also developed rapidly. However, with the surge in electricity demand, the contradiction between power supply and demand has become increasingly prominent. In order to alleviate the supply and demand contradiction, achieve efficient and precise investment by power grid companies, and improve the operating efficiency and economy of the power system, there is an urgent need to further improve the time-of-use pricing mechanism. As an important foundation and key link of the time-of-use pricing strategy, the division of time periods directly affects the implementation effect of the time-of-use pricing strategy. Existing methods for dividing time periods for time-of-use pricing mainly include empirical analysis, factor analysis, and membership function methods. However, the implementation of these technical solutions still relies on human experience, resulting in inaccurate time period division and making it difficult to achieve a relatively reasonable time period division.

[0003] Therefore, how to achieve a more accurate and reasonable allocation of time-of-use electricity pricing is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0004] The purpose of this application is to provide a time-of-use pricing method for dividing time periods, which can achieve a more accurate and reasonable division of time-of-use pricing. Another purpose of this application is to provide a time-of-use pricing device, electronic device, and computer-readable storage medium, all of which have the above-mentioned beneficial effects.

[0005] Firstly, this application provides a method for dividing time-of-use electricity pricing periods, including:

[0006] Obtain power load samples, which include power load data collected at preset time intervals within a preset duration;

[0007] Determine the initial clustering parameters and the preset number of clusters for the Gaussian mixture model, wherein the preset number of clusters is the number of time periods to be divided for the preset duration;

[0008] Based on the power load sample and the preset number of clusters, the initial clustering parameters are estimated to obtain the optimal clustering parameters;

[0009] The time period to which each of the power load data collection points belongs is determined based on the optimal clustering parameters.

[0010] Optionally, based on the power load samples and the preset clustering number, parameter estimation is performed on the initial clustering parameters to obtain the optimal clustering parameters, including:

[0011] Based on the power load samples and the preset number of clusters, the initial clustering parameters are iteratively trained using the expectation-maximization algorithm to obtain the optimal clustering parameters.

[0012] Optionally, based on the power load samples and the preset number of clusters, the initial clustering parameters are iteratively trained using the expectation-maximization algorithm to obtain the optimal clustering parameters, including:

[0013] The posterior probability of each of the power load data in each cluster is calculated using the initial clustering parameters.

[0014] The initial clustering parameters are updated using the posterior probabilities to obtain the updated clustering parameters;

[0015] The updated clustering parameters are used as the initial clustering parameters, and the step of calculating the posterior probability of each power load data under each cluster using the initial clustering parameters is returned for iterative training until the preset iteration conditions are met.

[0016] The updated clustering parameters under the preset iteration conditions are taken as the optimal clustering parameters.

[0017] Optionally, the initial clustering parameters include the weights, expectations, and covariance matrices of each cluster. The posterior probability of each power load data point in each cluster is calculated using the initial clustering parameters, including:

[0018] The initial clustering parameters and the power load data are calculated using a probability calculation formula to obtain the posterior probability of each power load data point under each cluster; wherein the probability calculation formula is:

[0019]

[0020] Wherein, the γ s,i p(x) represents the posterior probability of the i-th power load data in the s-th cluster. i |β s Let x represent the probability density function of the i-th power load data in the s-th cluster. i Let be the i-th power load data, S be the preset cluster size, and α be the number of clusters. s Let be the weight of the s-th cluster.

[0021] Optionally, the initial clustering parameters are updated using the posterior probabilities to obtain updated clustering parameters, including:

[0022] The updated clustering parameters are obtained by calculating the posterior probabilities and the initial clustering parameters using a parameter update formula; wherein the parameter update formula is:

[0023]

[0024]

[0025]

[0026] Where r is the number of power load data, α s Let μ be the weight of the s-th cluster. s Let ∑ be the expectation under the s-th cluster. s Let be the covariance matrix of the s-th cluster.

[0027] Optionally, the preset iteration condition is that the number of iterations reaches a preset number of iterations.

[0028] Optionally, determining the time period to which each of the power load data collection time points belongs based on the optimal clustering parameters includes:

[0029] Under the optimal clustering parameters, for each of the power load data, the maximum posterior probability is determined among all posterior probabilities;

[0030] The cluster corresponding to the maximum posterior probability is used as the time period to which the collection time point of the power load data belongs.

[0031] Secondly, this application also discloses a time-of-use pricing time-sharing device, comprising:

[0032] The acquisition module is used to acquire power load samples, which include power load data collected at preset time intervals within a preset duration.

[0033] The determination module is used to determine the initial clustering parameters and the preset number of clusters for the Gaussian mixture model, wherein the preset number of clusters is the number of time periods to be divided for the preset duration;

[0034] The calculation module is used to estimate the initial clustering parameters based on the power load sample and the preset number of clusters, and obtain the optimal clustering parameters.

[0035] The segmentation module is used to determine the segmented time period to which each of the power load data collection time points belongs based on the optimal clustering parameters.

[0036] Thirdly, this application also discloses an electronic device, comprising:

[0037] Memory, used to store computer programs;

[0038] A processor, configured to execute the computer program to implement any of the time-of-use pricing time-sharing methods described above.

[0039] Fourthly, this application also discloses a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of any of the time-of-use pricing time-sharing methods described above.

[0040] This application provides a time-of-use pricing method for dividing time periods, comprising: acquiring power load samples, the power load samples including power load data collected at preset time intervals within a preset duration; determining initial clustering parameters and a preset number of clusters for a Gaussian mixture model, the preset number of clusters being the number of time periods to be divided for the preset duration; estimating the initial clustering parameters based on the power load samples and the preset number of clusters to obtain optimal clustering parameters; and determining the time period to which the collection time point corresponding to each power load data belongs based on the optimal clustering parameters.

[0041] Applying the technical solution provided in this application, power load data is first collected at preset time intervals within a preset duration to obtain power load samples for model training. This facilitates the acquisition of a Gaussian training model with optimal clustering parameters. During model training, after obtaining the initial clustering parameters and preset number of clusters for the Gaussian mixture model, the initial clustering parameters can be updated using the power load samples and preset number of clusters to obtain the optimal clustering parameters, thereby achieving the training of the Gaussian mixture model. Consequently, the optimal clustering parameters of the Gaussian mixture model can be used to achieve time-of-use pricing. It is evident that, based on this implementation method, the time-of-use pricing is achieved using the Gaussian mixture model clustering algorithm, effectively eliminating reliance on manual intervention and further realizing more accurate time-of-use pricing time-of-use division, ensuring the rationality of the time-of-use pricing time-of-use division.

[0042] The time-of-use pricing device, electronic device, and computer-readable storage medium provided in this application also have the above-mentioned technical effects, and will not be described in detail here. Attached Figure Description

[0043] To more clearly illustrate the technical solutions in the prior art and the embodiments of this application, the accompanying drawings used in the description of the prior art and the embodiments of this application will be briefly introduced below. Of course, the accompanying drawings described below with respect to the embodiments of this application are only a part of the embodiments in this application. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort, and such other drawings also fall within the protection scope of this application.

[0044] Figure 1 A flowchart illustrating a time-of-use pricing method provided in this application;

[0045] Figure 2 A flowchart illustrating another time-of-use pricing method provided in this application;

[0046] Figure 3 This application provides a daily load data variation curve.

[0047] Figure 4 A comparison chart of peak-hour electricity consumption percentage under different time period division methods provided in this application;

[0048] Figure 5 A schematic diagram of a time-of-use pricing device provided in this application;

[0049] Figure 6 This is a schematic diagram of the structure of an electronic device provided in this application. Detailed Implementation

[0050] The core of this application is to provide a time-of-use pricing method for dividing time periods, which can achieve a more accurate and reasonable division of time-of-use pricing. Another core aspect of this application is to provide a time-of-use pricing device, electronic device, and computer-readable storage medium, all of which have the aforementioned beneficial effects.

[0051] To provide a clearer and more complete description of the technical solutions in the embodiments of this application, the technical solutions in the embodiments of this application will be described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0052] This application provides a method for dividing time-of-use electricity pricing periods.

[0053] Please refer to Figure 1 , Figure 1This is a flowchart illustrating a time-of-use pricing method provided in this application. The time-of-use pricing method may include the following steps S101 to S104.

[0054] S101: Obtain power load samples, which include power load data collected at preset time intervals within a preset duration;

[0055] This step aims to acquire electricity load samples, which are used for training a Gaussian mixture model to obtain a Gaussian mixture model with optimal clustering parameters. This optimal clustering parameter facilitates the time-of-use pricing based on these parameters. The electricity load sample includes multiple electricity load data points, which can be collected at preset time intervals within a preset duration. For example, when dividing daily time-of-use pricing into time periods, the preset duration can be 24 hours, and the preset time interval can be 1 hour. This means that electricity load data can be collected every hour within 24 hours, resulting in an electricity load sample containing 24 data points. Of course, this is merely an example; the specific values ​​of the preset duration and time interval do not affect the implementation of this technical solution. They can be set by technical personnel according to actual needs, and this application does not impose any limitations on them.

[0056] S102: Determine the initial clustering parameters and preset number of clusters for the Gaussian mixture model. The preset number of clusters is the number of time periods for a preset duration.

[0057] This step aims to determine the initial clustering parameters and the preset number of clusters for the Gaussian mixture model. It should be noted that this application aims to implement time-of-use pricing based on the Gaussian mixture model clustering algorithm. Therefore, the Gaussian mixture model can be trained first to obtain a Gaussian mixture model with optimal clustering parameters. These optimal clustering parameters are obtained through continuous optimization based on the initial clustering parameters; therefore, the Gaussian mixture model parameters need to be initialized first to obtain the initial clustering parameters. Furthermore, the clustering algorithm also involves the number of clusters. Therefore, the number of clusters also needs to be preset. In this application, the preset number of clusters is the number of time-of-use divisions for a preset duration. For example, the time-of-use pricing for a 24-hour day can be divided into time periods to obtain peak, average, and off-peak electricity periods. In this case, the preset number of clusters is set to 3.

[0058] Among them, the Gaussian mixture model (GMM) clustering algorithm aims to classify members based on probability. It belongs to "soft classification," meaning it doesn't explicitly categorize members into specific classes but rather provides the probability of each member belonging to each class. This method contains significantly more information than "hard classification" clustering methods like K-means, resulting in better clustering performance. Furthermore, GMM clustering is better at uncovering the correlations between various attributes. A GMM is obtained by linearly combining multiple Gaussian distribution functions. If an r-dimensional sample dataset is divided into S classes, then the Gaussian mixture model (composed of a mixture of S Gaussian distributions) is defined as follows:

[0059]

[0060]

[0061]

[0062] Where P(x) is the probability density function of the GMM, p(x|β) s Let α represent the probability density function of the s-th cluster. s Let μ be the weight of the s-th cluster. s Let ∑ be the expectation under the s-th cluster. s Let be the covariance matrix of the s-th cluster.

[0063] S103: Based on the power load samples and the preset number of clusters, estimate the initial clustering parameters to obtain the optimal clustering parameters;

[0064] This step aims to calculate the optimal clustering parameters. As mentioned above, the power load samples are used to train the Gaussian mixture model to obtain a Gaussian mixture model with optimal clustering parameters. Therefore, after obtaining the power load samples, the preset number of clusters, and the initial clustering parameters, the initial clustering parameters can be estimated using the power load samples and the preset number of clusters. That is, parameter updates are achieved through model training to obtain the optimal clustering parameters. Obviously, the Gaussian mixture model corresponding to these optimal clustering parameters is the optimal model finally obtained through training.

[0065] S104: Determine the time period to which each power load data collection time point belongs based on the optimal clustering parameters.

[0066] This step aims to achieve time-of-use pricing based on optimal clustering parameters. As mentioned above, each power load data in the power load sample is collected at preset time intervals within a preset duration. Each power load data corresponds to a collection time point. Based on this, the time-of-use pricing is divided into time periods for each collection time point within the preset duration, determining the time period to which each collection time point belongs, such as determining whether each collection time point belongs to peak electricity consumption, flat electricity consumption, or valley electricity consumption.

[0067] As can be seen, the time-of-use pricing method provided in this application first collects power load data at preset time intervals within a preset duration to obtain power load samples for model training. This facilitates the acquisition of a Gaussian training model with optimal clustering parameters through model training. During model training, after obtaining the initial clustering parameters and preset number of clusters for the Gaussian mixture model, the initial clustering parameters can be updated using the power load samples and preset number of clusters to obtain the optimal clustering parameters, thereby achieving the training of the Gaussian mixture model. Consequently, the optimal clustering parameters of the Gaussian mixture model can be used to achieve time-of-use pricing. Therefore, based on this implementation method, the time-of-use pricing is achieved using the Gaussian mixture model clustering algorithm, effectively eliminating reliance on manual intervention and further achieving more accurate time-of-use pricing time-of-use division, ensuring the rationality of the time-of-use pricing time-of-use division.

[0068] Based on the above embodiments:

[0069] In one embodiment of this application, the above-mentioned estimation of initial clustering parameters based on power load samples and a preset number of clusters to obtain optimal clustering parameters may include: iteratively training the initial clustering parameters using the expectation-maximization algorithm (EM) based on power load samples and a preset number of clusters to obtain optimal clustering parameters.

[0070] This application provides a method for estimating initial clustering parameters based on power load samples and a preset number of clusters to obtain optimal clustering parameters, which can be implemented based on the expectation-maximization algorithm. The basic idea of ​​the expectation-maximization algorithm is to solve for the maximum likelihood estimate of the model distribution parameters and iterate the model parameters repeatedly until the likelihood function converges, thus completing the parameter estimation. Furthermore, since solving the likelihood estimation function for GMM sample data is relatively complex, the logarithm of the likelihood function is often taken, as shown in the following formula:

[0071]

[0072] To achieve accurate parameter estimation, the EM algorithm mainly performs iterative estimation through the following two steps: E-Step (Expectation Step) and M-Step (Maximization Step). In E-Step, the probability of each data point being generated by each Gaussian distribution is calculated using the initial values ​​of α, μ, and ∑ or the values ​​obtained from the previous iteration. In M-Step, the model parameters are solved and updated using the values ​​obtained from E-Step.

[0073] In one embodiment of this application, the above-mentioned method of iteratively training the initial clustering parameters using the expectation-maximization algorithm based on the power load samples and the preset number of clusters to obtain the optimal clustering parameters may include the following steps:

[0074] The posterior probability of each power load data under each cluster is calculated using the initial clustering parameters.

[0075] The initial clustering parameters are updated using the posterior probabilities to obtain the updated clustering parameters;

[0076] The updated clustering parameters are used as the initial clustering parameters. The steps of calculating the posterior probability of each power load data under each cluster using the initial clustering parameters are returned for iterative training until the preset iteration conditions are met.

[0077] The updated clustering parameters under the preset iteration conditions are taken as the optimal clustering parameters.

[0078] This application provides a method for calculating optimal clustering parameters based on the expectation-maximization algorithm. First, for each power load data point in the power load sample, the posterior probability of that data point in each cluster is calculated using initial clustering parameters. Then, the initial clustering parameters are trained and updated using these posterior probabilities to obtain updated clustering parameters. Further, these updated clustering parameters are used as initial clustering parameters to return to the posterior probability calculation step, thus achieving iterative training until a preset iteration condition is met. The updated clustering parameters obtained under the preset iteration condition are then the final optimal clustering parameters.

[0079] The preset iteration conditions can be set by technicians according to actual needs. In one possible implementation, the preset iteration condition can be that the number of iterations reaches a preset number of iterations. Of course, this setting is only one implementation provided by the embodiments of this application and is not the only one. For example, the preset iteration condition can also be set to the convergence of the training function.

[0080] In one embodiment of this application, the initial clustering parameters may include the weights, expectations, and covariance matrices of each cluster. The aforementioned calculation of the posterior probability of each power load data point within each cluster using the initial clustering parameters may include: calculating the posterior probability of each power load data point within each cluster using a probability calculation formula; wherein the probability calculation formula is:

[0081]

[0082] Where, γ s,i p(x) represents the posterior probability of the i-th power load data in the s-th cluster. i |β s Let x represent the probability density function of the i-th electricity load data in the s-th cluster. i Let S be the number of predefined clusters for the i-th power load data, and α be the number of clusters. s Let be the weight of the s-th cluster.

[0083] This application provides a method for calculating the posterior probability of power load data under clustering, which can be calculated based on the above formula.

[0084] In one embodiment of this application, the above-mentioned updating of the initial clustering parameters using each posterior probability to obtain updated clustering parameters includes: calculating the updated clustering parameters using each posterior probability and the initial clustering parameters using a parameter update formula; wherein, the parameter update formula is:

[0085]

[0086]

[0087]

[0088] Where r is the number of power load data, α s Let μ be the weight of the s-th cluster. s Let ∑ be the expectation under the s-th cluster. s Let be the covariance matrix of the s-th cluster.

[0089] This application provides a method for updating clustering parameters, which can be calculated based on the above formula.

[0090] In one embodiment of this application, determining the time period to which each power load data collection point belongs based on the optimal clustering parameters may include the following steps:

[0091] Under the optimal clustering parameters, for each power load data, determine the maximum posterior probability among all posterior probabilities;

[0092] The cluster corresponding to the maximum posterior probability is used as the time period to which the collection time point of the power load data belongs.

[0093] This application provides a method for dividing time-of-use electricity pricing into time periods based on optimal clustering parameters. It is understood that under optimal clustering parameters, for each power load data point, a preset number of posterior probabilities can be obtained. Therefore, the maximum posterior probability can be selected from all corresponding posterior probabilities, and the cluster corresponding to this maximum posterior probability is determined. This cluster then represents the time period to which the corresponding power load data's collection time point belongs. For example, when the cluster corresponding to the maximum posterior probability is peak electricity consumption period, and the corresponding power load data's collection time point is 9:00, then 9:00 belongs to the peak electricity consumption period.

[0094] Based on the above embodiments, this application provides another method for dividing time-of-use electricity pricing periods.

[0095] Please refer to Figure 2 , Figure 2 The following is a flowchart illustrating another time-of-use pricing method provided in this application. The implementation process of this time-of-use pricing method is as follows:

[0096] First, collect a sample set D of r-dimensional daily load data. r =(x1,x2,...,x r (For example: if a data point is collected every hour, then there are 24 time periods in a day, r = 24), and the number of clusters S is given (for example: if a day is divided into three time periods, namely peak, flat and valley, then S = 3), that is, GMM is a mixture of S Gaussian distributions.

[0097] Furthermore, given random initial parameters α0, μ0, and ∑0 for the model, the posterior probability γ of each data point generated by the s-th Gaussian distribution is calculated using the E-Step method. s,i :

[0098]

[0099] Where, γ s,i p(x) represents the posterior probability of the i-th power load data in the s-th cluster. i |β s Let x represent the probability density function of the i-th electricity load data in the s-th cluster. i Let S be the number of predefined clusters for the i-th power load data, and α be the number of clusters. s Let be the weight of the s-th cluster.

[0100] Then, based on the values ​​obtained in the E-Step, the model parameters are updated using the M-Step method:

[0101]

[0102]

[0103]

[0104] Where r is the number of power load data, α s Let μ be the weight of the s-th cluster. s Let ∑ be the expectation under the s-th cluster. s Let be the covariance matrix of the s-th cluster.

[0105] Repeat the E-Step and M-Step steps until the stopping conditions are met (e.g., reaching the maximum number of iterations or the convergence of the log-likelihood function) to complete the estimation of the GMM parameters.

[0106] Finally, when the GMM parameters are known, based on the data sample x i (i = 1, 2, ..., r) can be generated by various Gaussian distributions for their posterior probabilities. The cluster with the highest posterior probability is selected as the cluster to which x belongs for that time period. i The cluster label for (i = 1, 2, ..., r) is σ. i (The cluster label to which the i-th sample belongs):

[0107] σ i =argmaxγ s,i ,i=1,2,...,r; s∈{1,2,...,S};

[0108] Then, based on the cluster label σ i Data sample set D r =(x1,x2,...,x r Divide into S clusters T all ={T1,T2,...,T S}, then the resulting S clusters T all This involves dividing the time period into S categories, with each category containing the time periods belonging to that category, thus completing the division of time-of-use electricity pricing time periods.

[0109] Based on this, in the embodiments of this application, taking typical daily load data of a domestic power system as an example, the effectiveness of the proposed time-of-use pricing method based on GMM clustering algorithm is verified.

[0110] Please refer to Figure 3 , Figure 3The daily load data variation curve provided in this application is set with a cluster size K=3, dividing the 24 time periods of the day into peak periods, normal periods, and valley periods. Currently, the time period division in the peak-valley time-of-use electricity pricing policy implemented in this region is as follows: 11:00-17:00 and 20:00-22:00 are peak periods; 8:00-11:00, 17:00-20:00, and 22:00-24:00 are normal periods; and 0:00-8:00 is the valley period. However, due to... Figure 3 It is evident that the current time-sharing method fails to fully reflect the peak-valley characteristics of the load in this region and contains numerous inconsistencies. For example, at 10:00, the load shows a clear upward trend and a high load value, but the current time-sharing system classifies this period as a normal period; at 21:00, the load shows a clear downward trend, but the current time-sharing system classifies this period as a peak period. Therefore, it is necessary to adjust the time-sharing system for this region.

[0111] To fully demonstrate the effectiveness of the time-of-use pricing method based on the GMM clustering algorithm in this application, methods M0-M3 were set up for comparative analysis:

[0112] M0: Current time period division method;

[0113] M1: Time-of-use pricing method based on membership function;

[0114] M2: A time-of-use pricing method based on K-means clustering algorithm;

[0115] M3: A time-of-use pricing method based on the GMM clustering algorithm.

[0116] In M1-M3: M1 is the traditional time-of-use pricing method, which is simple and efficient, but is greatly affected by human factors; M2 is the classic clustering method commonly used for time-of-use pricing, which is simple in principle and easy to implement, but has weak adaptability; M3 is the method proposed in this application.

[0117] Referring to this comparative analysis method, the percentage of electricity generated during peak periods under different time period divisions was statistically analyzed to determine the total electricity generated during those periods. Based on these percentages, and through comprehensive analysis, the rationality of the time period division results was evaluated. The time period division results for M0-M3 are shown in Table 1. The percentage of electricity generated during peak periods under the M0-M3 time period division is as follows: Figure 4 As shown in Table 1, this application provides a time-of-use pricing (TOU) allocation result for M0-M3. Figure 4 This application provides a comparison chart of peak-hour electricity consumption percentages under different time-period division methods.

[0118] Table 1 Results of Time-of-Use Pricing for M0-M3

[0119]

[0120]

[0121] Based on the time-sharing results of M0-M3 obtained in Table 1, by comparing the proposed time-of-use pricing time-sharing methods M3 with M1 and M2, it can be concluded that: in the time-sharing results of M1, 9:00 was not allocated to the peak period, but... Figure 3 It can be seen that the load is in a significant upward trend at 9:00 and the load value is relatively high, only 189.14MW away from the peak (12:00), which is about 86.88% of the peak. Therefore, if this period is classified as a valley period, a new load peak may occur at this point after users transfer their loads, which is not conducive to peak shaving and valley filling. In the time period division results of M2, 13:00 is not classified as a peak period. Although the load at this point is in a downward trend, its load value is relatively high, only 46.82MW away from the peak (12:00), which is about 96.75% of the peak. If this period is classified as a normal period, there may be less transferable load or a new load peak may occur at this point after users complete the load transfer, which will make it impossible to achieve peak shaving and valley filling effectively.

[0122] In addition, refer to Figure 4 By comparing the peak-hour electricity consumption percentages of M3 and M0-M2, we can see that M3 increased by 5.64%, 4.27%, and 4.75% compared to M0-M2, respectively. Under a scientific time-of-use pricing policy, users will actively shift their load from periods with high electricity prices to periods with low electricity prices. Therefore, the higher the peak-hour electricity consumption percentage, the more load users can transfer, which is beneficial for users to transfer more load from peak hours to normal or valley hours, thus achieving peak shaving and valley filling of the load curve to a greater extent.

[0123] As can be seen, the time-of-use pricing method provided in this application first collects power load data at preset time intervals within a preset duration to obtain power load samples for model training. This facilitates the acquisition of a Gaussian training model with optimal clustering parameters through model training. During model training, after obtaining the initial clustering parameters and preset number of clusters for the Gaussian mixture model, the initial clustering parameters can be updated using the power load samples and preset number of clusters to obtain the optimal clustering parameters, thereby achieving the training of the Gaussian mixture model. Consequently, the optimal clustering parameters of the Gaussian mixture model can be used to achieve time-of-use pricing. Therefore, based on this implementation method, the time-of-use pricing is achieved using the Gaussian mixture model clustering algorithm, effectively eliminating reliance on manual intervention and further achieving more accurate time-of-use pricing time-of-use division, ensuring the rationality of the time-of-use pricing time-of-use division.

[0124] This application provides a time-of-use electricity pricing time-division device.

[0125] Please refer to Figure 5 , Figure 5 This application provides a schematic diagram of a time-of-use pricing time-sharing device, which may include:

[0126] Module 1 is used to acquire power load samples, which include power load data collected at preset time intervals within a preset duration.

[0127] Module 2 is used to determine the initial clustering parameters and the preset number of clusters for the Gaussian mixture model. The preset number of clusters is the number of time periods for a preset duration.

[0128] Calculation module 3 is used to estimate the initial clustering parameters based on the power load sample and the preset number of clusters, and obtain the optimal clustering parameters;

[0129] The segmentation module 4 is used to determine the time period to which the collection time point corresponding to each power load data belongs based on the optimal clustering parameters.

[0130] As can be seen, the time-of-use pricing time-sharing device provided in this application first collects power load data at preset time intervals within a preset duration to obtain power load samples for model training. This facilitates the acquisition of a Gaussian training model with optimal clustering parameters through model training. During model training, after obtaining the initial clustering parameters and preset number of clusters of the Gaussian mixture model, the initial clustering parameters can be updated using the power load samples and preset number of clusters to obtain the optimal clustering parameters, thereby achieving the training of the Gaussian mixture model. Thus, the optimal clustering parameters of the Gaussian mixture model can be used to achieve time-of-use pricing time-sharing. It is evident that based on this implementation method, the time-of-use pricing time-sharing is achieved based on the Gaussian mixture model clustering algorithm, effectively eliminating reliance on manual intervention and further achieving more accurate time-of-use pricing time-sharing, ensuring the rationality of time-of-use pricing time-sharing.

[0131] In one embodiment of this application, the computing module 3 may include:

[0132] The iterative unit is used to iteratively train the initial clustering parameters using the expectation-maximization algorithm based on the power load samples and the preset number of clusters, in order to obtain the optimal clustering parameters.

[0133] In one embodiment of this application, the above-mentioned iterative unit may include:

[0134] The computational subunit is used to calculate the posterior probability of each power load data in each cluster using the initial clustering parameters.

[0135] The updated sub-unit is used to update the initial clustering parameters using each posterior probability to obtain the updated clustering parameters;

[0136] The iterative subunit is used to take the updated clustering parameters as the initial clustering parameters, return the steps of calculating the posterior probability of each power load data under each cluster using the initial clustering parameters, and perform iterative training until the preset iteration conditions are met.

[0137] Determine the sub-unit, which is used to take the updated clustering parameters under the preset iteration conditions as the optimal clustering parameters.

[0138] In one embodiment of this application, the initial clustering parameters include the weights, expectations, and covariance matrices of each cluster. The aforementioned calculation subunit can be specifically used to calculate the initial clustering parameters and power load data using a probability calculation formula to obtain the posterior probability of each power load data point in each cluster; wherein the probability calculation formula is:

[0139]

[0140] Where, γ s,i p(x) represents the posterior probability of the i-th power load data in the s-th cluster. i |β s Let x represent the probability density function of the i-th electricity load data in the s-th cluster. i Let S be the number of predefined clusters for the i-th power load data, and α be the number of clusters. s Let be the weight of the s-th cluster.

[0141] In one embodiment of this application, the aforementioned update subunit can be specifically used to calculate each posterior probability and the initial clustering parameters using a parameter update formula to obtain updated clustering parameters; wherein, the parameter update formula is:

[0142]

[0143]

[0144]

[0145] Where r is the number of power load data, α s Let μ be the weight of the s-th cluster. s Let ∑ be the expectation under the s-th cluster. s Let be the covariance matrix of the s-th cluster.

[0146] In one embodiment of this application, the preset iteration condition can be that the number of iterations reaches a preset number of iterations.

[0147] In one embodiment of this application, the division module 4 may include:

[0148] The first determining unit is used to determine the maximum posterior probability among all posterior probabilities for each power load data under the optimal clustering parameters.

[0149] The second determining unit is used to determine the cluster corresponding to the maximum posterior probability as the time period to which the collection time point of the power load data belongs.

[0150] For a description of the apparatus provided in the embodiments of this application, please refer to the above method embodiments; further details will not be repeated here.

[0151] This application provides an electronic device.

[0152] Please refer to Figure 6 , Figure 6 This application provides a schematic diagram of the structure of an electronic device, which may include:

[0153] Memory, used to store computer programs;

[0154] A processor, used to execute computer programs, can implement the steps of any of the time-of-use pricing time-sharing methods described above.

[0155] like Figure 6 The diagram shows the structural composition of an electronic device, which may include a processor 10, a memory 11, a communication interface 12, and a communication bus 13. The processor 10, memory 11, and communication interface 12 all communicate with each other through the communication bus 13.

[0156] In this embodiment, the processor 10 may be a central processing unit (CPU), an application-specific integrated circuit, a digital signal processor, a field-programmable gate array, or other programmable logic devices.

[0157] The processor 10 can call the program stored in the memory 11. Specifically, the processor 10 can execute the operations in the embodiment of the time-of-use pricing method.

[0158] The memory 11 is used to store one or more programs. The programs may include program code, which includes computer operation instructions. In this embodiment, the memory 11 stores at least a program for implementing the following functions:

[0159] Acquire power load samples, which include power load data collected at preset time intervals within a preset duration;

[0160] Determine the initial clustering parameters and the preset number of clusters for the Gaussian mixture model. The preset number of clusters is the number of time periods to be divided for a preset duration.

[0161] Based on the power load samples and the preset number of clusters, the initial clustering parameters are estimated to obtain the optimal clustering parameters;

[0162] The time period to which each power load data collection point belongs is determined based on the optimal clustering parameters.

[0163] In one possible implementation, the memory 11 may include a program storage area and a data storage area, wherein the program storage area may store the operating system and applications required for at least one function; and the data storage area may store data created during use.

[0164] In addition, memory 11 may include high-speed random access memory, and may also include non-volatile memory, such as at least one disk storage device or other volatile solid-state storage device.

[0165] Communication interface 12 can be an interface for the communication module, used to connect with other devices or systems.

[0166] Of course, it should be noted that, Figure 6 The structure shown does not constitute a limitation on the electronic device in the embodiments of this application. In practical applications, the electronic device may include more than Figure 6 More or fewer components as shown, or combinations of certain components.

[0167] This application provides a computer-readable storage medium.

[0168] The computer-readable storage medium provided in this application embodiment stores a computer program, which, when executed by a processor, can implement the steps of any of the time-of-use pricing time-sharing methods described above.

[0169] The computer-readable storage medium may include various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0170] For a description of the computer-readable storage medium provided in the embodiments of this application, please refer to the above method embodiments; further details will not be repeated here.

[0171] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to the method section.

[0172] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0173] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented directly by hardware, a software module executed by a processor, or a combination of both. The software module can be located in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.

[0174] The technical solutions provided in this application have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the embodiments above are only for the purpose of helping to understand the methods and core ideas of this application. It should be noted that those skilled in the art can make several improvements and modifications to this application without departing from the principles of this application, and these improvements and modifications also fall within the protection scope of this application.

Claims

1. A method for dividing time-of-use electricity pricing periods, characterized in that, include: Obtain power load samples, which include power load data collected at preset time intervals within a preset duration; Determine the initial clustering parameters and the preset number of clusters for the Gaussian mixture model, wherein the preset number of clusters is the number of time periods to be divided for the preset duration; Based on the power load sample and the preset number of clusters, the initial clustering parameters are estimated to obtain the optimal clustering parameters; The time period to which each of the power load data collection time points belongs is determined based on the optimal clustering parameters. Specifically, the process of estimating the initial clustering parameters based on the power load samples and the preset number of clusters to obtain the optimal clustering parameters includes: iteratively training the initial clustering parameters using the expectation-maximization algorithm based on the power load samples and the preset number of clusters to obtain the optimal clustering parameters; The process involves iteratively training the initial clustering parameters using the expectation-maximization algorithm based on the power load samples and the preset number of clusters to obtain the optimal clustering parameters. This includes: calculating the posterior probability of each power load data point in each cluster using the initial clustering parameters; updating the initial clustering parameters using the posterior probabilities to obtain updated clustering parameters; using the updated clustering parameters as the initial clustering parameters, and returning to the step of calculating the posterior probability of each power load data point in each cluster using the initial clustering parameters for iterative training until a preset iteration condition is met; and using the updated clustering parameters under the preset iteration condition as the optimal clustering parameters. The initial clustering parameters include the weights, expectations, and covariance matrices of each cluster. Calculating the posterior probability of each power load data point in each cluster using the initial clustering parameters includes: calculating the posterior probability of each power load data point in each cluster using a probability calculation formula; wherein the probability calculation formula is: ； Among them, the Indicates the first The power load data in the first Posterior probabilities under each cluster Indicates the first The power load data in the first The probability density function under each cluster, For the first Electricity load data, The preset number of clusters, For the first Weights under each cluster.

2. The time-of-use pricing method according to claim 1, characterized in that, The initial clustering parameters are updated using the posterior probabilities to obtain updated clustering parameters, including: The updated clustering parameters are obtained by calculating the posterior probabilities and the initial clustering parameters using a parameter update formula; wherein the parameter update formula is: ； ； ； ； in, The quantity of the power load data. For the first Weights under each cluster For the first Expectations under each cluster, For the first The covariance matrix under each cluster.

3. The time-of-use pricing method according to claim 1, characterized in that, The preset iteration condition is that the number of iterations reaches a preset number of iterations.

4. The time-of-use pricing method according to claim 1, characterized in that, The time period to which each of the power load data collection time points belongs is determined based on the optimal clustering parameters, including: Under the optimal clustering parameters, for each of the power load data, the maximum posterior probability is determined among all posterior probabilities; The cluster corresponding to the maximum posterior probability is used as the time period to which the collection time point of the power load data belongs.

5. A time-of-use pricing device, characterized in that, include: The acquisition module is used to acquire power load samples, which include power load data collected at preset time intervals within a preset duration. The determination module is used to determine the initial clustering parameters and the preset number of clusters for the Gaussian mixture model, wherein the preset number of clusters is the number of time periods to be divided for the preset duration; The calculation module is used to estimate the initial clustering parameters based on the power load sample and the preset number of clusters, and obtain the optimal clustering parameters. The segmentation module is used to determine the segmented time period to which the collection time point corresponding to each of the power load data belongs based on the optimal clustering parameters. The calculation module includes an iterative unit, which is used to iteratively train the initial clustering parameters using the expectation-maximization algorithm based on the power load sample and the preset number of clusters to obtain the optimal clustering parameters; The iterative unit includes: The calculation subunit is used to calculate the posterior probability of each of the power load data in each cluster using the initial clustering parameters; An update subunit is used to update the initial clustering parameters using the posterior probabilities of each parameter, thereby obtaining updated clustering parameters. An iterative subunit is used to take the updated clustering parameters as the initial clustering parameters and return to the step of calculating the posterior probability of each power load data under each cluster using the initial clustering parameters for iterative training until the preset iteration conditions are met. A subunit is determined, which is used to take the updated clustering parameters under the preset iteration conditions as the optimal clustering parameters; The initial clustering parameters include the weights, expectations, and covariance matrices of each cluster. The calculation subunit is specifically used to calculate the initial clustering parameters and the power load data using a probability calculation formula to obtain the posterior probability of each power load data point under each cluster. The probability calculation formula is as follows: ； Among them, the Indicates the first The power load data in the first Posterior probabilities under each cluster Indicates the first The power load data in the first The probability density function under each cluster, For the first Electricity load data, The preset number of clusters, For the first Weights under each cluster.

6. An electronic device, characterized in that, include: Memory, used to store computer programs; A processor, configured to execute the computer program to implement the steps of the time-of-use pricing method as described in any one of claims 1 to 4.

7. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the time-of-use pricing method as described in any one of claims 1 to 4.

Images